U.S. patent number 9,736,625 [Application Number 15/384,506] was granted by the patent office on 2017-08-15 for enhanced wireless communication for medical devices.
The grantee listed for this patent is Eko Devices, Inc.. Invention is credited to Tyler Crouch, Eugene Gershtein, Connor Landgraf.
United States Patent |
9,736,625 |
Landgraf , et al. |
August 15, 2017 |
Enhanced wireless communication for medical devices
Abstract
Methods and apparatuses for wireless communication between
medical devices are provided. In some embodiments, commodity low
power, low bandwidth communication protocols may be utilized to
simultaneously convey multiple signals with high fidelity and
reliability. For example, cardiac sound data and ECG data may be
compressed using a common ADPCM component and inserted into a
common BLE packet structure. Command-control data may also be
inserted. Where required command-control data reporting frequency
is less than the packet frequency, header bits may be utilized to
convey multiple types of command-control data in a given packet
byte position. Rolling packet sequence values may be inserted into
the common packet structure, for use by receiving devices to
identify link integrity failures.
Inventors: |
Landgraf; Connor (San
Francisco, CA), Gershtein; Eugene (Redwood City, CA),
Crouch; Tyler (San Francisco, CA) |
Applicant: |
Name |
City |
State |
Country |
Type |
Eko Devices, Inc. |
Berkley |
N/A |
CA |
|
|
Family
ID: |
59562608 |
Appl.
No.: |
15/384,506 |
Filed: |
December 20, 2016 |
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H04W
4/38 (20180201); A61B 5/318 (20210101); A61B
5/316 (20210101); A61B 7/00 (20130101); A61B
5/0006 (20130101); H04W 4/80 (20180201); H04W
4/70 (20180201); A61B 5/002 (20130101); G16H
40/67 (20180101); Y02D 30/70 (20200801); G16H
40/63 (20180101) |
Current International
Class: |
A61B
5/00 (20060101); A61B 5/0402 (20060101); A61B
5/04 (20060101); H04W 4/00 (20090101) |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Bilodeau; David
Attorney, Agent or Firm: Bertoglio; Brad
Claims
What is claimed is:
1. A method for transmitting cardiac data from a wireless sensor to
a host device, the method comprising: digitizing cardiac sound data
and ECG data received at the wireless sensor; filtering the
digitized cardiac sound data and ECG data; compressing the cardiac
sound data and the ECG data using an adaptive differential
compression component; combining the compressed cardiac sound data
and compressed ECG data into a common packet structure; and
transmitting the common packet structure from the wireless sensor
to the host device via a Bluetooth Low Energy communications
link.
2. The method of claim 1, in which the step of compressing the
cardiac sound data and ECG data comprises applying both the cardiac
sound data and the ECG data to a common adaptive differential pulse
code modulation encoder.
3. The method of claim 2, in which the ECG data has a sample rate
of approximately 500 Hz.
4. The method of claim 3, in which the cardiac sound data has a
sample rate of approximately 4 kHz.
5. The method of claim 4, in which the step of filtering the
digitized cardiac sound data comprises applying a digital lowpass
filter to attenuate frequency components above 2 kHz.
6. The method of claim 1, in which the step of combining the
compressed cardiac sound data and compressed ECG data into a common
packet structure further comprises inserting command-control data
into the common packet structure.
7. The method of claim 6, in which the command-control data has a
reporting frequency lower than the packet frequency, wherein the
step of inserting command-control data into the common packet
structure comprises inserting a command-control comprising a
command-control value and a header value, the header value
indicative of a content type with which the command-control value
is associated.
8. The method of claim 1, in which the step of combining the
compressed cardiac sound data and compressed ECG data into a common
packet structure further comprises inserting a rolling packet
sequence value into the common packet structure.
9. The method of claim 8, in which the rolling packet sequence
value is four bits in length.
10. The method of claim 8, further comprising: receiving by the
host device sequential data packets having non-sequential rolling
packet sequence values; and displaying a warning indicia on a host
device user interface.
Description
TECHNICAL FIELD
The present disclosure relates to medical devices utilizing
wireless electronic communications. More specifically, this
disclosure relates to methods and apparatuses for enhancing
wireless communications in medical device applications, such as
wireless cardiac sensors.
BACKGROUND
Use of wireless communications techniques for electronic devices is
becoming increasingly popular. Wireless devices provide convenience
and ease of use. Bluetooth has become particularly prevalent as a
wireless communications protocol. It provides versatile mechanisms
for transmitting digital signals over short distances with very low
power consumption. Bluetooth has become a ubiquitous standard
amongst mobile phones, tablet computers, personal computers,
wireless headphones, automobiles, and a wide variety of other
device types. As a result, Bluetooth devices are readily
interoperable with other electronic devices. Meanwhile, high
production volumes result in ready availability and relatively low
cost for transceiver chipsets and circuit boards, further
reinforcing the widespread adoption of the standard.
Bluetooth Low Energy ("BLE") is a subprotocol defined within the
Bluetooth 4.0 protocol, that enables highly energy-efficient
transfer of data between a client device (e.g. a sensor) and a
server device (e.g. a mobile phone or personal computer). BLE can
be particularly valuable for battery-operated devices, for which
minimizing power consumption may be critical.
While the prevalence of Bluetooth and power-efficiency of BLE
provide many advantages, some device types, particularly in the
context of medical instrumentation, give rise to communication
requirements that may not be well-satisfied by standard Bluetooth
implementations. For example, many types of instrumentation may
require transmission of multiple signal types, which would
traditionally be conveyed by multiple wires or multiple wireless
radios. However, consumer electronic devices may be limited in the
number of radios provided, while sensors with multiple radios may
require greater power consumption, resulting in larger batteries
and/or worse battery life. Meanwhile, BLE bandwidth limitations may
impact sensor performance. For example, while humans can typically
perceive sounds ranging from about 20 Hz to about 20 kHz, BLE as a
protocol does not have enough bandwidth to transmit the entirety of
the human audio spectrum, due to small packet size and slow packet
speed. Traditional Bluetooth and BLE implementations may be
particularly disadvantageous or limiting for medical devices such
as wireless cardiac devices.
SUMMARY
Improved implementations of BLE-based wireless communication
protocols can provide high levels of performance in wireless
medical device applications, while still enabling use of commodity
Bluetooth transceiver hardware and commodity host electronic
devices.
In some embodiments, a method is provided for transmitting cardiac
data from a wireless sensor to a host device. Cardiac sound data
and ECG data are received at a wireless sensor, such as via onboard
transducers digitizing audio and electrical signals sensed on a
patient. The cardiac sound and ECG data can be filtered, such as
via application of a digital lowpass filter to cardiac sound data
to attenuate frequency components above approximately 2 kHz. The
cardiac sound data and ECG data are compressed, such as through
application of the data to an adaptive differential compression
component. In some embodiments, a common adaptive differential
compression component can be applied to both the cardiac sound data
and the ECG data. The compressed cardiac sound data and compressed
ECG data can be combined into a common packet structure, and
transmitted from the wireless sensor to the host device.
The common packet structure may also include command-control data.
In embodiments where the packet frequency is greater than the
required frequency of command-control data reporting,
command-control data may include a header bit indicating one of
multiple command-control data content types with which an
associated command-control value is associated--thereby reducing
the number of bits that must be allocated to command-control data
within the packet structure.
The common packet structure may also include mechanisms to identify
wireless communication link integrity problems. A packet sequence
value, such as a rolling four-bit value, can be inserted into each
packet by the transmitting device, such as a cardiac sensor. The
receiving device, e.g. the host device, can decode the packet
sequence value towards ensuring that sequentially-received packets
have sequential packet sequence values. In the event that the
receiving device identifies a gap in rolling packet sequence
values, the receiving device may determine the existence of a
failure of the wireless communication link integrity. Such a
failure may then be conveyed to a user via, e.g., displaying a
warning indicia on a host device user interface.
BRIEF DESCRIPTION OF THE FIGURES
FIG. 1 is a schematic block diagram of a medical instrumentation
environment including a wireless cardiac sensor and host
device.
FIG. 2 is a schematic block diagram of a wireless packet
structure.
FIG. 3 is a schematic block diagram of a cardiac signal processing
chain.
DETAILED DESCRIPTION OF THE DRAWINGS
While this invention is susceptible to embodiment in many different
forms, there are shown in the drawings and will be described in
detail herein several specific embodiments, with the understanding
that the present disclosure is to be considered as an
exemplification of the principles of the invention to enable any
person skilled in the art to make and use the invention, and is not
intended to limit the invention to the embodiments illustrated.
Techniques are described that can be used to effectively transmit
medical device data, particularly heart diagnostic data, via a
low-power, low-bandwidth wireless communications protocol such as
Bluetooth Low Energy. Several techniques described hereinbelow can
be applied individually or in combination.
FIG. 1 illustrates a typical operating environment in which
embodiments can be employed. Cardiac sensor 100 is a wireless heart
monitor capable of detecting multiple types of diagnostic data,
including heart sounds and ECG electrical recordings. Sensor 100
includes microprocessor 102 for processing and storing data from
transducers 103 into memory 104. Sensor transducers 103 can include
audio transducer 103A, for auscultation such as recording of heart
sounds, and ECG transducer 103B, for monitoring of cardiac
electrical activity. Bluetooth transceiver 105 is in operable
communication with processor 102 in order to convey data to and
from remote electronic devices, such as host device 120. Battery
106 is a rechargeable battery supplying power to sensor 100. In
order to maximize the duration between required charges, and
minimize the size, weight and expense of sensor 100, sensor 100 is
designed for low power consumption during operation.
Sensor 100 communicates via wireless data connection 110 with host
device 120. Host device 120 may preferably be a standard, commodity
mobile wireless computing device, such as a smartphone (e.g. Apple
iPhone.TM.), tablet computer (e.g. Apple iPad.TM.), or laptop
computer. Host device 120 includes microprocessor 122 for
processing and storing data. Bluetooth transceiver 123 enables
wireless communication between processor 122 and external devices,
such as sensor 100. Host device 120 further includes user interface
components 124 (such as a touchscreen), memory 125 for data
storage, and battery 126. While illustrated as a mobile device in
the embodiment of FIG. 1, in other embodiments, host device 120
could alternatively be selected from amongst other types of
computing devices having a Bluetooth transceiver, such as a
personal computer or a central sensor monitoring station.
The BLE protocol may be desirable for implementation of wireless
communications link 110, in order to minimize energy consumption
during operation and therefore extend the battery life of sensor
100 and host device 120. However, BLE, as commonly implemented,
presents significant limitations in a wireless cardiac sensor
environment. One such limitation is bandwidth. Common mobile
devices 120 have limitations in packet rate utilizing the BLE
protocol for communications link 110. For example, some mobile
phones may have a theoretical minimum packet interval at which one
BLE packet can be accepted every 5 milliseconds. Exacerbating this
limitation is a need in medical applications for high data
integrity and reliability. In such embodiments, it may not be
desirable to potentially sacrifice data integrity and link
reliability by requiring data transmission at or near theoretical
maximum packet rates. While decreasing packet rate may provide
better packet interval operating margin, bandwidth constraints are
even more limiting. With some common consumer mobile devices, it
has been found that reliable BLE communications can be maintained
sending packets at 8 ms intervals.
BLE also imposes packet size constraints. Moreover, regardless of
protocol constraints on packet size, it may be further desirable to
reduce packet size in order to reduce power consumption. Meanwhile,
in order to implement an effective wireless cardiac sensor
providing both auscultation and ECG data, packets will preferably
accommodate multiple data streams, such as heart sound audio data
from audio transducer 103A, ECG data from transducer 103B, and
command-and-control data associated with the operation of cardiac
sensor 100 and its interaction with host device 120. Packet
efficiency may be critical to use of BLE in such environments.
FIG. 2 illustrates an optimized BLE packet structure that may be
utilized in communications from cardiac sensor 100 to host device
120. The packet structure of FIG. 2 is optimized to convey multiple
types of medical instrument and control data via a relatively
low-bandwidth and low-power BLE communication link that can be
reliably received by standard smartphones, tablets or other
consumer electronic devices. Specifically, the packet structure of
FIG. 2 conveys heart sounds, ECG data and command/control data
simultaneously, with clinical fidelity, within a single BLE packet,
using one standard BLE radio set.
Each packet 200 in FIG. 2 is preferably formed having a byte length
provided for by BLE standards, and packet intervals preferably
compatible with commodity BLE chipsets and computing devices. Such
a data structure may provide an effective bitrate of approximately
20 kbps.
Packet 200 includes header bytes 210, command and control bytes
220, and cardiac data 230. In the illustrated embodiment, cardiac
data 230 includes audio payload 232 and ECG payload 234. Audio
payload 232 is utilized for transmitting heart sound data recorded
by audio transducer 103A. FIG. 3 illustrates a schematic
representation of cardiac signal processing components within
cardiac sensor 100, which operate to generate data conveyed in the
BLE packet structure of FIG. 2. Audio sensor 300 converts an audio
signal, such as cardiac auscultation, into an analog electronic
signal. Analog-to-digital converter (ADC) 310 samples the output of
sensor 300 and generates a digital data stream 311. ADC 310
initially samples acoustic heart sound signals at an approximately
4 kHz sample rate, with 16-bit samples, yielding a 64 kbps audio
stream. Audio compression is applied by adaptive differential
pulse-code modulation (ADPCM) encoder 330 to yield a 4-bit audio
stream 332 at a 4 kHz rate (i.e. one 4-bit sample each 0.25 ms).
Therefore, with an 8 ms packet interval, each packet 200 includes
audio payload 232 having 32 4-bit audio samples.
Digital filters 320 can be applied to the output 311 of ADC 310
prior to ADPCM encoder 330 in order to reduce artifacts and
distortion during the ADPCM compression process. In particularly,
filters 320 will include strong low-pass filters to eliminate or
drastically attenuate high frequency components above the 2 kHZ
range. It has been determined that frequency range limitations
imposed by aggressive pre-filtering of cardiac auscultation sounds
before ADPCM compression is preferable for purposes of human
medical diagnostics, as compared to less aggressive filtering
accompanied by potential introduction of compression noise and
artifacts by ADPCM encoder 330.
Another advantage of the packet structure of FIG. 2, particularly
given limitations on packet interval in common smartphones and
other mobile devices that may be utilized as host device 120, is
that it combines heart sound and ECG data within a single BLE
packet. FIG. 3 further illustrates a schematic representation of an
ECG data pipeline that may be implemented on cardiac sensor 100. In
use, ECG sensors 340 are connected to a patient, and output
electrical signals 341 indicative of a patient's cardiac electrical
activity.
Cardiac electrical signals 341 are sampled by analog-to-digital
converter 350. In an exemplary embodiment, ADC 350 may generate
16-bit samples at a 500 Hz sampling rate. This yields a digital ECG
data stream 351 having a data rate of 8 kbps, to which filter 360
may be applied. Utilizing an 8 ms BLE packet interval, ECG data
stream 351 would therefore require 8 bytes within each BLE packet.
However, given the amount of packet 200 allocated to cardiac audio
data, as described above, it may be desirable to compress the ECG
data stream, provided the compression can be achieved without
material negative impact on the ECG data fidelity.
It has been determined that the same ADPCM encoder 330 used to
encode cardiac audio data, can also be effectively utilized to
reduce ECG data bandwidth without significant negative impact on
the ECG signal fidelity via strategic specification of sample rate.
By selecting a 500 Hz sample rate, measurement differentials
between adjacent samples in a typical digitized ECG signal are such
that the ECG data stream may be effectively encoded by ADPCM
encoder 330 to yield an encoded ECG data stream 334 that reduces
the size of ECG payload 234.
In some embodiments, audio sensor 300 and ADC 310 can be
implemented within audio transducer 103A, ECG sensors 340 and ADC
350 can be implemented within ECG transducer 103B, with filter 320,
filter 360 and encoder 330 being implemented by processor 102. In
other embodiments, the elements of FIG. 3 can be distributed
differently amongst components such as audio transducer 103A, ECG
transducer 103B, processor 102, custom ASICs, GPUs, or other
components.
Bandwidth-efficient conveyance of command and/or control data
(sometimes referred to as command-control data) may also be
important in wireless cardiac sensor and other medical device
applications. For command-control data of a nature that the
acceptable reporting frequency is less than the packet frequency,
it may be desirable for sequential packets to transmit different
command-control data content types within the same packet bit
positions. A header bit or bits may be utilized to indicate which
of multiple types of command-control data is conveyed within
associated packet bit positions.
For example, in the context of a wireless cardiac sensor
transmitting at an 8 ms packet interval, it may not be necessary to
transmit certain command-control data, such as volume level or
battery level, at 8 ms intervals. Longer intervals may be
sufficient, while still ensuring users perceive a high level of
responsiveness. Thus, in the packet structure of FIG. 2, bits
within header 210 can be utilized to convey one of multiple content
types of command-control data. For example, a header bit may be
utilized to indicate whether the data within command and control
data 220 reflects a volume level or battery level. Depending on the
number of bits required for sufficient command-control data value
granularity, and the desired frequency of command-control data
conveyance, in other embodiments, multiple header bits can be
utilized to enable greater numbers of command-control data content
types to be conveyed within a given packet byte position. For
example, in another embodiment, two bits may be used to specify one
of four different command-control data content types, with
associated bit positions conveying an associated value. In some
embodiments, header bits may be conveyed in different byte
positions from associated command and control values within packet
200; in other embodiments, header bits and associated command and
control values may be conveyed within the same byte position of
packet 200, thereby intermixing header data 210 and command and
control data 220.
Another important aspect of wireless communications in some medical
applications is verifying link integrity. For high risk data such
as heart sound and ECG data, it may be desirable for devices to
rapidly and reliably alert the user when a data transmission
quality problem arises. By effectively identifying data
transmission issues, a user can promptly remedy equipment problems
and ensure that anomalous results are attributed to instrumentation
error rather than the patient being monitored. However, traditional
BLE protocols do not provide mechanisms to determine when packets
are dropped.
Therefore, the packet of FIG. 2 preferably includes a link
integrity verification mechanism integrated within the packet
structure. Predetermined bits within header 210 can be allocated to
a rolling packet sequence indicator. When transmitted by cardiac
sensor 100, processor 102 constructs consecutive packets to
increment through a rolling multi-bit packet sequence value. The
receiving device 120 can then decode the packet sequence value to
verify that consecutive packets are received with sequentially
incrementing packet sequence values. In the event that packet
sequence values are not sequential in adjacent packets, receiving
device 120 can determine that the integrity of link 110 has been
compromised, and alert a user to the issue by, e.g., displaying an
appropriate warning indicia on user interface 124. The embodiment
of FIG. 2 utilizes a rolling packet sequence value that is four
bits in length, which in some embodiments may be an optimal
tradeoff between minimizing failures to identify link integrity
problems (for which longer sequence values are better), and
minimizing power consumption and bandwidth attributed to the link
integrity verification function (for which shorter sequence values
are better).
The previous description of the disclosed embodiments is provided
to enable any person skilled in the art to make or use the
invention disclosed herein. Various modifications to these
embodiments will be readily apparent to those skilled in the art,
and the generic principles defined herein may be applied to other
embodiments without departing from the spirit or scope of the
disclosure. Thus, the present disclosure is not intended to be
limited to the embodiments shown herein but is to be accorded the
widest scope consistent with the principles and novel features
disclosed herein. All references cited herein are expressly
incorporated by reference.
* * * * *